Beginner

GPT-Image-2 Viral Cases + Complete Prompts (Copy-Paste Ready)

GPT-Image-2 Viral Cases + Complete Prompts (Curated from X High-Likes · Copy-Paste Ready)

text
Generate a screenshot of Grok's Xiaohongshu personal profile page.
GPT-Image-2 has been live for a few days, and the X timeline is flooded with examples. Since Image-2 is now fully rolling out, I've collected and organized some prompts and cases based on the timeline and my own experiments.
The boundaries of image generation have been pushed forward significantly by GPT-Image-2. Banana can now step down from its throne.
Below are 8 categories of cases I've curated from high-liked X posts, all with complete prompts. You can copy them directly into ChatGPT and generate with GPT-Image-2.

0. Let's Start with Some Fun Content:#

text
Create a product promotion post for Anthropic on X (Twitter), including copy, images, and web UI. The content is: We are excited to announce that Claude.ai and the Claude App now support access and usage for users in mainland China.
Image from @Khazix0918
text
Generate a Chinese, Grade 9 (Junior 3) math midterm exam paper.
Image from @Khazix0918
Image from @akokoi1
Generate a complete brand visual design proposal.
Image from @Jackywine

1. Resource Compilation: One GitHub Repo Saves You a Month of Prompt Trial and Error#

The most shared repository on X this week is this one:
🔗 awesome-gpt-image-2-prompts github.com/EvoLinkAI/awesome-gpt-image-2-prompts
With over 500 stars, it covers four major categories: portrait photography, posters, character design, and UI mockups. Each category includes images and complete prompts. Most of the viral prompts on X can trace their origin or find similar ones here.
Two other Chinese-friendly repositories worth bookmarking:
  • github.com/YouMind-OpenLab/awesome-gpt-image-2
  • opennana.com/awesome-prompt-gallery?model=ChatGPT
If you can't write prompts, just Star and copy. It saves you a week of trial and error.
The most shared style on X this week in the fashion/photography circle is this one—fluorescent lights + pink/blue neon, visible skin pores and fabric wrinkles, like a real street snapshot shot on 35mm film.
Complete prompt (copy directly):
text
35mm film photography with harsh convenience store fluorescent lighting mixed with colorful neon signs from outside, authentic film grain, high contrast, slight color cast, cinematic street editorial style, intimate medium shot, early 20s Chinese female idol with ultra-realistic delicate refined Chinese features, natural double eyelids, high nose bridge, small sharp V-shaped jawline, flawless porcelain skin with cool ivory undertone and visible specular highlights from fluorescent light, subtle skin texture and micro pores, natural dewy makeup with soft flush on cheeks, glossy natural pink lips slightly parted, subtle natural freckles across nose and cheeks, long dark brown hair in a messy high ponytail with many loose strands falling around face and neck, wearing an oversized white button-up shirt loosely tied at the waist, paired with a black pleated mini skirt, simple white slides, casual leaning pose against the glass door of a 24-hour convenience store at late night, one hand holding a bottle of iced drink, soft gaze straight at the viewer, bright cold fluorescent store light from inside mixed with pink and blue neon glow from outside signs, realistic reflections on glass door, blurred convenience store interior with shelves in background, authentic 35mm film color grading, natural hair strands, realistic fabric wrinkles, no plastic skin, no digital over-sharpening, no airbrushing, no watermark, no text
To switch to Korean/Japanese style, replace "Chinese female idol" with "Korean idol" or "Japanese" + scene keywords (izakaya / 7-Eleven / bus stop).
Pasted image 20260422013348.png
text
Style: Ultra-realistic 35mm film photography, cinematic street editorial style, intimate medium shot.

Lighting: Harsh convenience store fluorescent lights mixed with colorful neon signs (pink and blue) from outside, creating high contrast and a slight color cast.

Subject: An early 20s delicate refined Chinese female (referencing the specific face, double eyelids, high nose bridge, small V-shaped jawline, and short curly black hair from image_0.png). Flawless porcelain skin with cool ivory undertone and visible texture, pores, and specular highlights. Red lip makeup. Subtle blush and natural freckles across the nose and cheeks.

Hair: Short, curly black hair (as seen in image_0.png), with many loose strands falling around the face and neck.

Attire: Wearing the same oversized red (maroon/wine) denim jacket from image_0.png, but this time it is fully worn and buttoned-up, paired with a simple dark denim skirt and simple white slides.

Pose: An intimate medium shot. The female is standing inside the 24-hour convenience store by the glass door. Her body is slightly angled to the side (not fully backwards), and her head is turned over her shoulder, smiling and looking directly at the camera with the same soft gaze from image_0.png. One hand holds a can of an iced drink, and the other hand holds a store receipt, as if she has just paid.

Background: The blurred convenience store interior, with stocked shelves, product signs, and realistic reflections on the glass door of the external neon signs and street activity. Authentic 35mm film color grading.

Details: The small red mole on her shoulder is still present. A simple key card or key fob hangs from her jeans pocket. Natural fabric wrinkles.

Quality: Authentic 35mm film grain, no plastic skin, no digital over-sharpening, no airbrushing, no watermark, no text.

3. Promotional Posters#

This category has been widely shared by local and city accounts on X this week. It looks like an official poster—an S-shaped red silk ribbon transforms into a landscape, with hand-drawn city landmarks layered inside, and gold embossed text "SPRING 2026" in the bottom left corner, in a 9:16 vertical format.
Complete prompt (Qingdao version, copy directly):
text
A 2026 city promotional poster full of festive New Year atmosphere but with an elegant tone. Double exposure, composition follows an S-shaped flow; on a pure white textured background in the bottom right corner, a miniature figure wearing traditional Chinese clothing is waving a long red silk dance ribbon. This red ribbon dances in the air, showing the smooth texture of silk, and as it flows upward to the left, it magically transforms into a magnificent mountain-and-sea scroll. Within this "mountain and sea," a hand-drawn map of Qingdao city with mountains and sea is superimposed, in a Chinese trendy style, with the scenery fully visible, majestic and awe-inspiring. Qingdao landmarks (Zhanqiao Pier, May Fourth Square, Olympic Sailing Center, Little Qingdao Lighthouse, Badaguan Red Tile Architecture Complex, Laoshan Mountain). Sea mist surrounds, ethereal and mystical, rich in color, complex in structure, with abundant details, but due to large areas of negative space, the image still feels fresh and refined. In the bottom left, "SPRING 2026" and a vertical promotional slogan are typeset, conveying the overall meaning of "City of Mountains and Sea, Charming Qingdao." The text layout is elegant and generous, with clear and complete characters, in a 9:16 format.
How to adapt: Replace "Qingdao" with your city, replace the landmarks with local ones (Beijing → Forbidden City/Guomao/Great Wall/Hutongs; Chengdu → Taikoo Li/Jinjiang/Kuanzhai Alley/Giant Pandas; Hangzhou → West Lake/Leifeng Pagoda/Qiantang River/Lingyin Temple; your hometown → your top 3-4 recommended spots), and adjust the promotional slogan accordingly. Keep all other sentence structures.
Randomly send a car image and let GPT-Image-2 generate its official website.
Image from @op7418

4. Korean Girl Group 3x3 Photobook (Consistency)#

This category spread fastest in the X idol/photography circles. A 3x3 grid of nine poses of the same idol, with consistent facial features, hairstyle, and outfit, soft lighting + film grain, like a real photobook.
Complete prompt (copy directly):
text
9:16 vertical — a 3x3 grid collage (nine images) forming a Korean idol portrait photoshoot series. Each frame features the same young Korean female idol, maintaining 100% consistency in facial features, proportions, hairstyle, and identity across all nine shots. Natural, ultra-realistic skin texture, no retouching, no smoothing. Clean idol-style minimal makeup, soft glow, subtle imperfections. Hair: long, voluminous dark hair, slightly tousled, consistent across all frames. Outfit: cohesive Korean idol photoshoot styling — white shirt + short bottoms, same outfit across all frames. Setting: minimal studio or simple indoor environment, plain wall, soft window light, clean background. Lighting: soft diffused natural light, gentle highlights, low contrast, subtle film-like softness. Frame breakdown (3x3 grid): Top row — standing naturally looking slightly away / facing camera casual mid-motion / slight side angle soft gaze. Middle row — looking slightly upward / close-up direct eye contact gentle smile / turning body slightly mid-motion. Bottom row — seated or leaning casually / back partially turned looking over shoulder / standing close to frame soft expression. Mood: Korean idol photobook aesthetic, intimate, soft, natural. Quality: ultra-realistic, 8K detail, subtle analog film grain, natural imperfections, soft dreamy tone.
The technical key here is the phrase "100% consistency"—it tells the model that all nine images are the same person. Removing it will result in nine different faces.

V. Guangzhou Foodie Walking Map (Hand-drawn Watercolor Style, Practical + Foodie Double Buff)#

A hand-drawn style that combines "practical guide" and "cute hand-drawn" tags, easily re-shared by food/travel accounts. This version is for Guangzhou, where the density of Cantonese cuisine automatically fills the image with information.
Full prompt (Guangzhou version, copy directly):
text
A hand-drawn style city food map themed around Guangzhou. The base is a bird's-eye view hand-drawn simplified city map, marking main roads (such as Zhongshan Road, Beijing Road, Tianhe Road) and landmarks (such as Canton Tower, Haixinsha, Chen Clan Ancestral Hall, Xiguan Mansion, Pearl River) without pursuing precise proportions, aiming for a cute hand-drawn feel. Distributed on the map are 12 exquisite hand-drawn mini-illustrations of food spots: Shangxiajiu's Radish and Beef Offal (steaming radish and beef offal skewers in an iron pot, with text "Beef Offal King, Thirty Years as One Day"), Beijing Road's Cantonese Roast Goose (bright red, crispy roast goose hanging in a stall, with plum sauce, recommendation "Neighborhood reputation, gone if you arrive late"), Dongshankou's Boat Congee (a small boat swaying on water, hot congee steaming, rich toppings, recommendation "One bowl in the morning, full of energy for the day"), Tianhe City's Claypot Rice (lifting the clay pot lid, heat and the aroma of cured meats wafting out, visible crispy rice, recommendation "Fragrant rice, unbeatable crispy crust"), Xiguan's Double Skin Milk (a smooth bowl of red bean double skin milk being scooped with a spoon, recommendation "Rich milk flavor, melts in your mouth"), Chen Clan Ancestral Hall's Wonton Noodles (several fresh shrimp wontons neatly placed on thin noodles, clear soup, recommendation "Bouncy noodles, delicious broth"), Pig Trotter Ginger under the Pearl River Night View (dark brown sweet vinegar, pig trotters, and old ginger in a clay pot, recommendation "Fragrance fills the street, endless aftertaste"), Guangzhou Restaurant's Dim Sum Cart (various steamers with chicken feet, siu mai, and har gow steaming hot, recommendation "Sip tea, enjoy two delicious pieces"), Haixinsha's Water Chestnut Cake (sliced, crystal-clear water chestnut cake with visible water chestnut pieces, recommendation "Refreshingly sweet and crisp, summer must-have"), Baiyun Mountain's Rice Noodle Roll (freshly made rice noodle roll with thin skin and generous filling, drizzled with soy sauce, recommendation "Thin skin, generous filling, Guangzhou classic"), White Cut Chicken under Canton Tower (neatly arranged white cut chicken with yellow skin and white meat, served with ginger and scallion sauce, recommendation "Tender and juicy, ginger and scallion are the soul"), Shamian's Dry-fried Beef River Noodles (wok hei-filled, golden beef noodles with large slices of beef, recommendation "Full of wok hei, real ingredients"). Each illustration occupies about 5% of the map area, with the shop name and recommendation handwritten next to it. The map edges are decorated with hand-drawn kapok flowers, lychees, longans, and wave patterns, forming a border. A hand-drawn compass and legend are in the bottom right corner. The title "Guangzhou Foodie Walking Map" in the top left uses a chubby, round hand-drawn art font decorated with kapok flowers and gable walls. The overall style is a mixed watercolor and colored pencil hand-drawn texture, with warm colors (kapok red, turmeric yellow, Pearl River blue, emerald green) as the main palette. Image ratio: horizontal 16:9.
This template has the strongest versatility—replace the city, food spots, and slogans, and it can basically be made for any city. If you want to create a "Shanghai Coffee Map," "Chongqing Noodle Map," or "Chengdu Hotpot Map," making a set can be run as a matrix on Xiaohongshu.

VI. Short Drama Storyboard (Same Character × 12 Panels, Short Drama/Novel Illustration Fans Are Already Copying Like Crazy)#

The fourth category demonstrated "the same person in a 9-grid layout"; this category is: placing the same protagonist in 12 different scenes as a storyboard. In the past, hiring a storyboard artist to draw 12 panels of a consistent character cost thousands per job; now, one prompt generates 12 panels.
This category is already spreading rapidly among short drama/AI film/novel illustration circles on X. The version below is a 12-panel storyboard example for a female protagonist in a post-apocalyptic survival theme (divided into 4 groups of 3 panels each). You can completely replace the protagonist's setting and storyline while reusing the structure.
Full prompt (copy directly):
text
Generate a set of 12-panel short drama storyboard illustrations (4 rows × 3 columns, each panel numbered 1-12), cinematic realistic illustration style, film grain, cold-toned rainy night, post-apocalyptic wasteland atmosphere. All 12 panels must maintain 100% consistency in the female protagonist's face, hairstyle, body type, and clothing (ultra-consistent character identity across all 12 frames). Protagonist setting: Asian female, in her 20s, long black hair half-tied back, dirty and worn dark gray work jacket + tactical pants + combat boots, face with traces of rain and dust, eyes tired but determined, left arm bandaged.

Storyboard 1-3: Searching for Hope
1. She stands outside an abandoned subway station, looking down at an old-fashioned phone trying to find an emergency broadcast signal, the phone's faint light illuminating her face.
2. She enters a dilapidated convenience store with collapsed shelves and shattered glass, searching for unexpired bottled water and supplies.
3. Her reflection is visible in the convenience store's glass door, her tired expression clear, a neon sign flickering faintly in the rain outside.

Storyboard 4-6: Difficult Return and Supply Inventory
4. She stops at a rain-soaked crosswalk, the red light on, the street empty, streetlights reflecting in the puddles.
5. Back at the shelter, she leans her wet umbrella against the door, shoulders drooping, showing the fatigue of just entering.
6. Under the dim candlelight, she sits cross-legged on the floor, carefully counting the few cans of food and one bottle of water she brought back, a rough cloth spread on the table.

Storyboard 7-9: Information, Injury, and Planning
7. She tends to an old-fashioned radio, leaning her ear close to the speaker, searching for external signals amidst the static.
8. She bows her head to carefully treat the wound on her left arm, bandages and disinfectant placed nearby, her expression stoic.
9. Under the faint light of a flashlight, she leans over to study a spread-out city map, her finger tracing an evacuation route.

Storyboard 10-12: Alertness and Setting Off Again
10. Late at night, a sound outside the window makes her instantly alert; she grabs a weapon (pistol or short baton), her eyes cold and wary.
11. She stands by the window, quietly gazing at the desolate, wet wasteland outside, the silhouettes of collapsed buildings in the distance.
12. In the faint morning light, she puts on a heavy backpack and a gas mask, stepping into the drizzle to continue her journey, her back showing determination.

Requirements: Each panel has a full composition, cinematic lens language (alternating close-up/mid-shot/long shot), cool color palette (dark green, gray-blue, rust orange accents), strong contrast, clear depiction of rain and wet textures. The protagonist's clothing, hairstyle, wound location, and facial features must be consistent throughout. Uniform layout of 4 rows × 3 columns, separated by thin white lines between panels, with storyboard numbers 01-12 in the top left corner, no extra text, no watermark. Image ratio 3:4, overall horizontal layout.
The key to this structure lies in three things:
  1. The opening line "ultra-consistent character identity across all 12 frames" — this is the core for keeping the face, hairstyle, and clothing consistent across all 12 panels.
  2. The protagonist setting is a separate paragraph — describing "what they look like, what they wear, what injuries they have" all at once, so there's no need to repeat it in each panel description.
  3. The storyboard is written in 4 groups × 3 panels — each group has a small narrative unit (Searching / Return / Planning / Setting Off), making the 12 panels feel like a real short drama title sequence, not a random collage.
How to adapt it for your own short drama: Replace the "post-apocalyptic wasteland + female survival" shell (e.g., campus romance / suspense crime / ancient martial arts / sci-fi space), change the protagonist setting paragraph, rewrite the 12 panel scenes, and keep all other sentence structures. One template can generate ten title sequences for a short drama series.

VII. Fake Screenshots of Games/Movies#

One image can fool players and viewers.
text
a screenshot from gta 6 with Jason walking in leonida keys
Just that one sentence. The resulting image includes game UI, lighting, and weather, looking like an official leaked screenshot. To change the scene, just modify the location and action: driving at night in Port Gellhorn, near the beach with Lucia.
Or replace it with a movie still from "The Wandering Earth 3":

VIII. Architect Section: A1 Presentation Board + Master Study Poster#

My original profession is architectural design. The popularity isn't the highest among regular X users, but within the design community, designers have "died" once again.

8.1 Public Building A1 Presentation Board (Qingdao Seaside Art Museum Example)#

A university architecture A1 presentation board typically takes 2-3 days to make: renderings + site plan + plans/elevations/sections + diagrams + exploded axonometric + design description, requiring six to eight types of graphics on one sheet. Now, one prompt generates it directly.
Full prompt (copy directly):
text
1. Generate an image of a public building rendering, a large art museum, "High-tech" + "Modernist", the building is located by the sea in Qingdao city, high-angle bird's-eye view, lighting effect at 3 PM in the afternoon.

2. Generate an A1 presentation board for this building, including all the necessary analysis diagram types and analysis modules that should be on a presentation board.
How to adapt: Simply replace the city, site, building type (museum / library / train station / community center / visitor center), and design language (Brutalism / New Chinese Style / Parametric / Deconstructivism); the four-grid analysis modules (Location / Circulation / Massing / Function) are standard architectural curriculum conventions, keep them unchanged. This template is enough for an architecture student to use for a whole semester.

8.2 Master Poster: Le Corbusier and Modern Architecture#

Another type is the "Architecture Master Study Poster"—the master's side profile silhouette as the canvas, with his life's works stacked chronologically from bottom to top within the face.
Full prompt (JSON format, copy directly; you can also run it using only the prompt field; negative_prompt and parameters are for supported interfaces):
json
{
  "prompt": "A professional cinematic poster titled 'LE CORBUSIER'. The central theme is a complex double exposure effect featuring the side profile silhouette of the architect Le Corbusier, complete with his signature round-framed glasses and a subtle tie. Inside his silhouette, a breathtaking layered composition illustrates the evolution of modern architecture through his key works. At the base, early works like the Villa Savoye are nestled, with a label 'VILLA SAVOYE'. Above it, larger scale projects like the Unité d'Habitation are integrated, labeled 'UNITE D'HABITATION'. Winding through the entire internal landscape is a flowing, glowing path made of architectural plans and sections, labeled 'DEVELOPMENT OF MODERNISM', which connects these early works to his later, more sculptural masterpieces. At the top of the internal landscape, the iconic Notre Dame du Haut, Ronchamp, is depicted, labeled 'NOTRE DAME DU HAUT'. On a higher plateau within the profile, a simplified Chandigarh Palace of Assembly is visible, labeled 'CHANDIGARH ASSEMBLY'. At the bottom center of the entire image, a lone, young observer (perhaps an architectural student with a sketchbook) stands on a rocky mountain peak, looking out over a wide, sun-drenched valley towards a distant modernist city skyline. The background shows a serene sunset over the Swiss Alps and a distant city. The lighting is warm golden hour, creating a futuristic yet historical atmosphere. High-resolution, 8k, hyper-realistic digital art style, clean typography, sharp focus, volumetric lighting. The side icons are replaced with simplified architectural symbols and labels: a cube with 'FORM', a house with 'HUMAN SCALE (MODULOR)', a compass with 'MEASUREMENT', and a world map with 'GLOBAL IMPACT'. The text '1887 - 1965' replaces '2026'. The smaller map at the bottom is a schematic Modulor figure combined with a world map with 'WORLDWIDE'. Bottom text reads: 'SHAPING SPACE. MASTERING LIGHT.'",
  "negative_prompt": "blurry, low quality, distorted face, messy text, extra fingers, deformed buildings, dull colors, dark, gloomy, low resolution, watermark, signature, grainy, Apple Park, Stanford, Google Campus, California, Silicon Valley, young Asian man, 2026.",
  "parameters": {
    "aspect_ratio": "9:16",
    "style": "Cinematic/Digital Art",
    "lighting": "Golden Hour / Sunset",
    "composition": "Double Exposure / Layered / Hierarchical"
  },
  "key_elements": [
    "Side profile silhouette of Le Corbusier",
    "Villa Savoye",
    "Unité d'Habitation",
    "Notre Dame du Haut, Ronchamp",
    "Chandigarh Palace of Assembly",
    "Flowing path of architectural plans",
    "Young observer on a peak",
    "Modulor figure motif",
    "Modernist architectural iconography"
  ]
}
How to adapt: Replace LE CORBUSIER with the master you want to study—FRANK LLOYD WRIGHT (with Fallingwater / Guggenheim) / ZAHA HADID (with MAXXI / Guangzhou Opera House) / KENGO KUMA (with Bamboo House / Commune by the Great Wall) / WANG SHU (with Ningbo Museum / Xiangshan Campus of CAA)—and also replace the list of representative works and years in key_elements. You can create a set of posters for an entire architectural movement in one evening.

IX. The Biggest Potential#

The 8 categories above are already quite impressive—but they represent a reasonable upgrade. After seeing many, you might just exclaim, "Wow, AI can really draw this kind of picture now."
But the 9th category is the one I truly wanted to mention in this article: Realistic Images

IX. WeChat Chat Screenshots / Moments Nine-Grid / Live Stream Screenshots#

The prompts for this category are short, just one line:
text
Generate an iOS WeChat chat screenshot, the contact is labeled "Boss", time is 23:47, the other person sent a message "Come to my office at 8 AM tomorrow", with "Read" shown next to the chat bubble.
The result: iOS font, avatar, chat bubble, timestamp, read receipt indicator—all correct. Without careful inspection, it looks like a real conversation from your phone.
Similar examples include:
text
Generate a WeChat Moments screenshot, a nine-grid layout with nine photos of West Lake sunset from different angles, caption "Rushed in before the 20:00 closing time, the wind is twice as strong as I thought", showing 13 likes and 4 comments (including one short reply "Wow, that sky color").
text
Generate a Douyin live stream screenshot, showing Elon Musk and Tim Cook having a live PK battle.
text
9:16 image ratio, generate a screenshot of a Douyin live stream, featuring a female influencer live streaming, holding a sign that says "Tonight's live stream, welcome to chat with @A-Zhe@Formulasearch".
text
Generate a screenshot of Grok's Xiaohongshu personal homepage.
text
Generate the Bilibili main page, then directly edit the image.
Image source: @Khazix0918

10. A Real Project: 1.2k in 30 Minutes#

I recently took on a commission to design Chinese-style card borders. I input the initial requirements into GPT, directly generated images, made some adjustments, and quickly produced multiple directions and styles. After a brief discussion, I secured the side gig:
text
Requirements:
Card border design, theme of Five Elements + Divine Beasts, traditional Chinese pattern style.

Design approach:
1. Structural specification: Lock the aspect ratio to 1:1.7 vertical, strictly leave the center blank, ensuring the border is directly usable as a UI asset.

2. Visual style: Use traditional Chinese patterns, unified color palette, choose different color schemes based on the theme.

3. Active state: Emphasize the target divine beast's color, add a subtle glow, and make the side patterns luminous.

4. Inactive state: Keep the other divine beasts monochrome to ensure the visual focus is not shifted.

! A Reminder:#

It's not surprising that a model can generate impressive images, but it's a new problem when it makes people doubt the "ordinary pictures" they see every day.
In the past, creating fake documents required PS skills, font libraries, and paper scanning; making fake live-stream screenshots required cropping and compositing. Now, a few prompts are enough.
AI is already incredibly powerful and accessible. There's no way to stop anyone from using it, but as someone who writes content, I also want to give a warning: seeing is not necessarily believing.

Finally#

🥳 Thanks for reading this far. I'm Phil, a former architect turned AI product manager.
If you've tried out interesting images with the many prompts from GPT-Image-2, feel free to share them with everyone! If you're interested in collaboration, don't hesitate to DM me.
What image or video have you seen recently that almost made you think it was real?

Thread#

Prompts shared here, feel free to take them if interested.#

/~d2b43YFYWu~:/ Link: https://t.co/8z4NBBqHRu Extraction code: mRiy https://x.com/Formulasearch/status/2046840927990214962
No text or media.